|
(11) | EP 1 017 088 B1 |
| (12) | EUROPEAN PATENT SPECIFICATION |
|
|
| (54) |
Selective silicidation process in non-volatile semiconductor memory devices Selektives Silizidierungsverfahren in nichtflüchtigen Halbleiterspeichern Procédé de silicidation sélective dans des mémoires semi-conductrices non-volatiles |
|
|
||||||||||||||||||||||||||||
| Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention). |
[0001] This invention relates to a selective silicidation process in the formation of non-volatile electronic memory devices integrated on a semiconductor substrate, said non-volatile memory devices comprising a plurality of active elements formed with gate and drive regions comprising at least one polysilicon layer.
[0002] The invention relates in particular, but not exclusively, to a selective silicidation process for the active elements of an EPROM or Flash EPROM cell, and the description to follow will consider this field of application for convenience of explanation.
Prior Art
[0003] As is well known, non-volatile memory devices integrated on a semiconductor substrate comprise:
- matrices of memory cells, wherein each cell comprises a floating gate MOS transistor; and
- control circuitry comprising fast-logic MOS transistors.
[0004] Each floating gate MOS transistor conventionally comprises a drain region and a source region which are formed in the semiconductor substrate and separated by a channel region. A floating gate electrode is formed over the substrate and isolated therefrom by a thin layer of gate oxide.
[0005] A control electrode is coupled capacitively to the floating gate electrode through a dielectric layer.
[0006] Each matrix of memory cells is organized into rows, known as the word lines, and columns, known as the bit lines. Cells belonging to the same word line have a common supply line which drives their respective control electrodes, and cells belonging to the same bit line have their drain terminals in common.
[0007] Where memory devices of very small size are to be formed, the interconnect lines between the gate electrodes, which may be on the order of 0.25µm, for example, comprise layers of a low resistivity material.
[0008] A first prior solution for providing such low resistivity material layers has been the use of composite material layers, known as silicides, which comprise silicon and a transition metal such as titanium, to cover those areas where resistivity is to be lowered.
[0009] The formation of a silicide layer over the active areas of MOS transistors comprises the following steps, following the formation of the transistor gate:
- implanting source and drain regions of the transistor;
- depositing a transition metal;
- subjecting the transition metal to a thermal process for selectively reacting it with the substrate surface and producing the silicide layer.
[0010] While being in several ways advantageous, this first prior solution still has certain drawbacks.
[0011] In fact, in applying the thermal process for reacting the transition metal layer with the substrate surface, thereby to produce the silicidation of the implanted source and drain regions, a surface layer of the substrate is consumed, and some of the dopant in the substrate leaks into the silicide layer. Accordingly, the silicide layer may become short-circuited to the substrate, thereby interfering with the normal operation of the cell.
[0012] The underlying technical problem of this invention is to provide a silicidation process for the gate electrodes of electronic devices, specifically memory devices, having such features that the formation of a silicide layer over the implanted regions in the semiconductor substrate can be prevented, thereby overcoming the drawbacks with which prior art silicidation processes have been beset.
Summary of the Invention
[0014] The solving idea behind this invention is that of providing an improved process flow for silicidizing electronic devices which comprises, before the silicidation step, a step of forming a dielectric layer to cover all the areas where the silicidation process may be critical and only leave uncovered those polysilicon portions where silicidation is instead desirable, without introducing any additional masking steps in the process.
[0015] Based on this solving idea, the technical problem is solved by a silicidation process as defined in Claim 1.
[0016] The features and advantages of a process according to the invention will be apparent from the following description of an embodiment thereof, given by way of non-limitative example with reference to the accompanying drawings.
Brief Description of the Drawings
[0017] In the drawings:
Figures 1 to 10 are respective enlarged vertical cross-section views of a portion of a semiconductor substrate as a silicidation process according to the invention progresses.
Detailed Description
[0018] Referring to the drawing views, an improved selective silicidation process for electronic devices 1 integrated on a semiconductor substrate 2, specifically memory devices of the EPROM and Flash EPROM types, will be described herein below.
[0019] Shown to an enlarged scale in the Figures are vertical cross-sections through portions, not necessarily adjacent one another, of a semiconductor substrate wherein the memory devices are formed in accordance with the invention.
[0020] Some process steps will not be discussed in detail in the description which follows, not to burden it with information of minor weight already available to the skilled persons in the art.
[0021] The first step in the process of fabricating EPROM or Flash EPROM memory devices conventionally comprises forming active areas 8 of all the active elements 3, such as floating gate MOS transistors 3b, of the matrix, and transistors 3a of the circuitry.
[0022] The definition of the active areas 8, isolated from one another by insulating regions consisting of a field oxide 9, is carried out conventionally.
[0023] Subsequently to defining the active areas 8, a gate oxide layer 10 may be grown for forming the transistors 3a, 3b.
[0025] A first resist mask for patterning the first polysilicon layer 11, referred to in the art as the POLY1 mask, is used to define the floating gate electrodes 4a of the memory cells.
[0026] By way of a conventional photolithographic step, the unprotected polysilicon layer by the POLY1 mask is etched away.
[0027] Upon removal of the POLY1 mask, an intermediate dielectric layer 12, e.g. of ONO (Oxide-Nitride-Oxide), is deposited.
[0028] A second resist mask, referred to in the art as the MATRIX mask, is then used for masking the portion of the semiconductor where the memory matrix is to be formed.
[0029] By a conventional photolithographic step, the first layer of polysilicon 11 and intermediate dielectric 12, being unprotected by the MATRIX mask, are etched away to expose the gate oxide layer 10 in the portion of the semiconductor where the circuitry is being formed.
[0031] At this step of the fabrication process, the control electrodes 4b and word lines WL of the cell matrix which will drive the control electrodes of the memory cells in the same column are defined.
[0032] A third resist mask, referred to in the art as the self-aligned etch mask and used for defining the second polysilicon layer 12, is then used for defining the word lines WL of the cell matrix.
[0033] The second layer of polysilicon 5, unmasked by the self-aligned etch mask, is etched away by a conventional photolithographic step.
[0035] Advantageously, before removing the self-aligned etch mask, a fourth mask, referred to in the art as the SAS mask, is provided.
[0036] By a conventional photolithographic step, the residual field oxide layer between the word lines WL is etched away as shown in Figure 7, this being followed by an ion implantation step to form the source and drain regions 13.
[0037] The process is continued with the deposition of a first dielectric layer 6 over the entire surface of the substrate. For example, this dielectric layer 6 could be of the TEOS type.
[0038] The thickness of the dielectric layer is illustratively within the range of 500Å to 3000Å, enough to isolate the transistor gate electrodes 4 and the word lines of the matrix.
[0039] A second dielectric layer 6a is deposited next to fill the gaps between the word lines of the matrix.
[0041] Alternatively, this second dielectric layer may be deposited, by a HDPCVD (High Density Plasma Chemical Vapor Deposition) technique, or a SACVD (Sub-Atmospheric Chemical Vapor Deposition) technique.
[0042] Nothing prevents from using a single dielectric filler layer 6 to fill the gaps between the word lines of the matrix.
[0043] At this point, the inventive process provides for a removal of the second dielectric layer 6b by a CMP (Chemical Mechanical Polishing) treatment to produce mechanically a planar surface.
[0045] The etch-back is carried out to advantage using no additional masks, and is continued until the surface of the second polysilicon layer 5 of the word lines and the surface of the second polysilicon layer in the circuitry become exposed.
[0046] In this way, the second dielectric layer is confined to just the gap regions between the word lines, and only the areas to be silicized are exposed.
[0047] At this point, a transition metal, such as titanium, is deposited and then processed thermally to cause it to react selectively with the uncovered areas by the dielectric layer and produce a silicide layer 7.
[0048] In summary, the process of this invention allows the resistance of the conductive polysilicon layers to be reduced, and silicidation of the source and drain regions of the memory matrix avoided, thereby preventing any dangerous bridging by the silicide between the source/drain regions and the first polysilicon layer or the substrate surface in the active areas, which would impair the proper operation of the matrix, without introducing any additional masking steps in the process flow.
- forming active areas of said active elements (3),
- forming a gate oxide layer (10) on the semiconductor substrate (2),
- forming a first conducting layer (11) on said gate oxide layer (10),
- patterning the first conducting layer (11) with a first resist mask,
- etching the first conducting layer (11) that is unprotected by the first resist mask to form floating gate electrodes (4a) of the memory matrix,
- forming an intermediate dielectric layer (12) on the entire surface of the semiconductor substrate (2),
- masking the memory matrix with a second resist mask,
- etching the intermediate dielectric layer (12) and the first conducting layer (11) that is unprotected by the second resist mask,
- forming a polysilicon layer (5) on the entire surface of the semiconductor substrate (2),
- patterning the polysilicon layer (5) with a third resist mask,
- etching the polysilicon layer (5) that is unprotected by the third resist mask to form control gate electrodes (4b) and word lines of the memory cells,
- forming source and drain regions (13) of the memory cells,
- depositing a dielectric layer (6) on the entire surface of the semiconductor substrate (2);
- removing said dielectric layer (6) to expose the polysilicon layer (5);
- depositing a layer of a transition metal (7);
- subjecting the transition metal layer to a thermal treatment for selectively reacting it with the polysilicon layers and producing a silicide layer (8) on said control gate electrodes (4b), said drive regions and said second polysilicon layer (5) in the circuitry.
- Bilden von aktiven Zonen der aktiven Elemente (3),
- Bilden einer Gate-Oxidschicht (10) auf dem Halbleitersubstrat (2),
- Bilden einer ersten leitenden Schicht (11) auf der Gate-Oxidschicht (10),
- Strukturieren der ersten leitenden Schicht (11) mit einer ersten Ätzschutzmaske,
- Ätzen der ersten leitenden Schicht (11), die von der ersten Ätzschutzmaske nicht geschützt ist, um Elektroden (4a), mit schwebendem Gate, der Speichermatrix zu bilden,
- Bilden einer dazwischen liegenden dielektrischen Schicht (12) auf der gesamten Oberfläche des Halbleitersubstrats (2),
- Maskieren der Speichermatrix mit einer zweiten Ätzschutzmaske,
- Ätzen der dazwischen liegenden dielektrischen Schicht (12) und der ersten leitenden Schicht (11), die von der zweiten Ätzschutzmaske nicht geschützt ist,
- Bilden einer Polysiliziumschicht (5) auf der gesamten Oberfläche des Halbleitersubstrats (2),
- Strukturieren der Polysiliziumschicht (5) mit einer dritten Ätzschutzmaske,
- Ätzen der Polysiliziumschicht (5), die von der dritten Ätzschutzmaske nicht geschützt ist, um Steuergate-Elektroden (4b) und Wortleitungen der Speicherzellen zu bilden,
- Bilden von Source- und Drainbereichen (13) der Speicherzellen,
- Abscheiden einer dielektrischen Schicht (6) auf der gesamten Oberfläche des Halbleitersubstrats (2);
- Entfernen der dielektrischen Schicht (6), um die Polysiliziumschicht (5) freizulegen;
- Abscheiden einer Schicht aus einem Übergangsmetall (7);
- Unterwerfen der Schicht aus Übergangsmetall einer Wärmebehandlung, um sie selektiv mit den Polysiliziumschichten zur Reaktion zu bringen, und Herstellen einer Silizidschicht (8) auf den Steuergate-Elektroden (4b), den Treiberbereichen und der zweiten Polysiliziumschicht (5) in der Schaltung.
- former des zones actives desdits éléments actifs (3),
- former une couche d'oxyde de grille (10) sur le substrat semiconducteur (2),
- former une première couche conductrice (11) sur ladite couche d'oxyde de grille (10),
- structurer la première couche conductrice (11) à l'aide d'un premier masque de resist,
- attaquer chimiquement de la première couche conductrice (11) qui n'est pas protégée par le premier masque de resist pour former des électrodes de grille flottantes (4a) de la matrice de mémoire,
- former une couche diélectrique intermédiaire (12) sur l'ensemble de la surface du substrat semiconducteur (2),
- masquer la matrice de mémoire avec un second masque de resist,
- attaquer chimiquement la couche diélectrique intermédiaire (12) et la première couche conductrice (11) qui n'est pas protégée par le second masque de resist,
- former une couche de polysilicium (5) sur l'ensemble de la surface du substrat semiconducteur (2),
- structurer la seconde couche de polysilicium (5) avec un troisième masque de resist,
- attaquer chimiquement la seconde couche semiconductrice (5) qui n'est pas protégée par le troisième masque de resist de manière à former des électrodes de grille de commande (4b) et les lignes de mots des cellules de mémoire,
- former des régions de source et de drain (13) des cellules de mémoire,
- déposer une couche diélectrique (6) sur l'ensemble de la surface du substrat semiconducteur (2),
- éliminer ladite couche diélectrique (6) pour exposer la couche de polysilicium (5);
- déposer une couche d'un métal de transition (7);
- soumettre la couche de métal de transition à un traitement thermique pour la faire réagir sélectivement avec les couches de polysilicium et produire une couche de siliciure (8) sur ladite électrode de grille de commande (4b), lesdites régions de commande et ladite seconde couche de polysilicium (5) dans le circuit.